When Usain competes, U.S. aid plummets.

 At Andrew Sullivan’s Daily Beast, Patrick Appel offers a few hypotheses about why Americans seem to care less about the killing of Sikhs than the killing of moviegoers, including the observation that the timing of the Milwaukee shootings so soon after the Batman massacre have left many pundits unwilling to talk further about gun control for fear of sounding redundant. Appel also hypothesizes that low levels of media coverage may be due to the Aurora killings haven taken place on a slow news day while the Milwaukee killings happened during the Olympics. Robert Wright in The Atlantic proposes a potentially complementary hypothesis: that the mass American public has cared less about Milwaukee than Aurora because of a sense that the Sikhs are outsiders while the theatergoers were representative.

Research by Thomas Eisensee and David Stromberg (ungated) suggests that at least two of these guesses may be right. Eisensee and Stromberg studied the effect of news coverage of more than 5,000 natural disasters on policymakers’ responses to see whether policy responses were driven by media coverage or policy rationales. Their study hinges on a fundamental truth about the media business: during large-scale events such as the Olympics, television networks, which have a fixed time budget (even a 24-hour-network can’t broadcast more than 24 hours a day), have less time to devote to unplanned events like disasters because of the time they spend on the scheduled spectacle. As Eisensee and Stromberg write,

If two equally newsworthy disasters occur, we would expect the disaster occurring when there is a great deal of other breaking news around would have a lower chance of being covered by the news than the disaster occur- ring when there is little other news around. This crowding out is probably particularly strong for television news broadcasts that are usually of a fixed length (half an hour for ABC, CBS and NBC, and one hour for CNN).

If policymakers’ responses are driven by some inherent logic of disasters and policy rationales, then the magnitude of their reactions should be unrelated to the availability of news coverage; if policymakers instead only act when the public is watching, then their responses to similar disasters should

Their results are startling–and dismaying. U.S. policymakers react to publicity, not severity.

First, media coverage is driven not by the severity of a disaster but by factors such as how people are killed and which people are killed:

News biases relief in favor of certain disaster types and regions: for every person killed in a volcano disaster, 40,000 people must die in a drought to reach the same probability of media coverage. Similarly, it requires 40 times as many killed in an African disaster to achieve the same expected media coverage as for a disaster in Eastern Europe of similar type and magnitude.

Second, the effects of media coverage are noticeable and substantively important:

We find that natural disasters are more likely to receive relief if they occur when the pressure for news time in the U.S. network news broadcasts is low. Quantitatively, disasters are, on average, around eight percent more likely to receive relief if they occur when news pressure takes on its highest values than when taking its lowest, and five percent less likely to receive relief during the Olympics than at other times. Using another metric, to have the same chance of receiving relief, the disaster occurring during the highest news pressure must have six times as many casualties as the disaster occurring when news pressure is at its lowest, all else equal.

It also turns out that the Olympics are the most important stories generating “news pressure”–the crowding out of foreign and disaster news–on the U.S. media, much more than the World Series, the Oscars, and the Super Bowl. Other sources of news, pressure, such as the O.J. Simpson murder trial verdict, perform similarly.

Eisensee and Stromberg conclude by asserting that although their story focused on domestic media coverage’s effects on U.S. policymakers’ efforts abroad, “it seems likely that the underlying mechanisms would be equally active for domestic policy.”